<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
seq-xls-r-fleurs_zu-run3-asr_xh-run3
This model is a fine-tuned version of lucas-meyer/xls-r-fleurs_zu-run3 on asr_xh. It achieves the following results:
- Val WER: 0.5206
- Test WER: 0.5318
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.4685 | 1.2 | 250 | 1.3087 | 0.9112 |
| 0.7027 | 2.39 | 500 | 0.4796 | 0.6514 |
| 0.455 | 3.59 | 750 | 0.4100 | 0.5497 |
| 0.3316 | 4.78 | 1000 | 0.3792 | 0.5202 |
| 0.2489 | 5.98 | 1250 | 0.4110 | 0.5374 |
| 0.2088 | 7.18 | 1500 | 0.4044 | 0.4694 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3
